Claude Code adds CLI auth management, faster startup, Windows ARM64

Highlight

New Auth Subcommands Bring Login Management to the CLI

You can now manage authentication directly from the terminal using claude auth login, claude auth status, and claude auth logout -- no need to navigate settings or restart a session. This is particularly useful in CI environments or when switching between accounts, giving you a clean, scriptable interface for credential management.

Worth Knowing

Claude Code Now Warns When You Launch It Inside Itself

If you accidentally try to start Claude Code from within an existing Claude Code session, you'll now get a clear warning instead of a confusing nested environment. This catches a common mistake when working in Claude-managed terminals.

Slash-Rename Now Generates Session Names Automatically

Calling /rename without any arguments now produces a name derived from your conversation context rather than leaving the field blank. You no longer need to think up a label -- just run /rename and Claude will summarize the session for you.

Native Windows ARM64 Binary Now Available

If you run Claude Code on a Windows ARM64 device, you'll now get a native binary instead of running under emulation. Expect better performance and lower overhead on Snapdragon-based Windows machines.

Prompt Cache Hit Rates Improved for Long Sessions

The current date has been moved out of the system prompt, which previously invalidated the prompt cache on every new day. Your longer or repeated sessions should now benefit from higher cache hit rates, reducing latency and token costs.

Startup Is Faster Thanks to Deferred Schema Loading

Zod schema construction is now deferred until it's actually needed, shaving time off cold starts. You'll notice Claude Code launches more quickly, particularly on slower machines or in CI pipelines.

Under the Hood

Terminal Rendering, File Handling, and Session Bugs Fixed

This release cycle resolves a VS Code terminal scroll regression, character loss at screen boundaries, blank lines in verbose transcript view, @-mention resolution with anchor fragments like @README.md#installation, FileReadTool blocking on FIFOs and large files, and ENAMETOOLONG errors for deeply-nested paths. Auth refresh errors and process hangs after session close are also fixed.

Hook Errors, MCP Crash, and Session Preview Display Fixed

Hook blocking errors (exit code 2) now correctly surface stderr output to you instead of silently failing. A crash when MCP tools return image content during streaming is resolved, and /resume session previews no longer show raw XML tags or interrupt messages as session titles.

From Anthropic

Claude Blog · 5 min read

Behind the model launch: What customers discovered testing Claude Opus 4.6 early

Four customer teams share how they stress-tested Opus 4.6 before release, including the technical edge cases they found and the feedback that shaped the final model. Worth reading if you're evaluating whether to move workloads to Opus 4.6.

Claude Blog · 5 min read

Behind the model launch: What customers discovered testing Claude Opus 4.6 early

Four customer teams share how they stress-tested Opus 4.6 before release, including the technical edge cases they found and the feedback that shaped the final model. Worth reading if you're evaluating whether to move workloads to Opus 4.6.

Claude Blog · 2 min read

Claude Enterprise, now available self-serve

Any organization can now purchase Claude Enterprise directly at claude.ai -- no sales call required. If your team has been waiting to get SSO, admin controls, and higher rate limits without a procurement cycle, this removes that friction.

Claude Blog · 2 min read

Claude Enterprise, now available self-serve

Any organization can now purchase Claude Enterprise directly at claude.ai -- no sales call required. If your team has been waiting to get SSO, admin controls, and higher rate limits without a procurement cycle, this removes that friction.

In the News

New York Times · 4 min read

Anthropic Pushes Its Valuation to $380 Billion With New Funding Round

Anthropic closed a $30 billion Series G round, more than doubling its valuation since September. The scale of investment signals continued confidence in Claude as a commercial platform and is likely to accelerate model development and infrastructure buildout.

New York Times · 4 min read

Anthropic Pushes Its Valuation to $380 Billion With New Funding Round

Anthropic closed a $30 billion Series G round, more than doubling its valuation since September. The scale of investment signals continued confidence in Claude as a commercial platform and is likely to accelerate model development and infrastructure buildout.

Reuters · 3 min read

Pentagon threatens to cut off Anthropic in AI safeguards dispute, Axios reports

The Pentagon is reportedly considering ending its contract with Anthropic over disagreements about usage restrictions -- specifically whether Claude can be used for mass surveillance and autonomous weapons. The outcome could influence how Anthropic shapes its usage policies going forward.

Reuters · 3 min read

Pentagon threatens to cut off Anthropic in AI safeguards dispute, Axios reports

The Pentagon is reportedly considering ending its contract with Anthropic over disagreements about usage restrictions -- specifically whether Claude can be used for mass surveillance and autonomous weapons. The outcome could influence how Anthropic shapes its usage policies going forward.

Wall Street Journal · 4 min read

Exclusive | Anthropic Takes Big Step in AI Race to Reshape College Coding Courses

Anthropic has partnered with hundreds of community and state colleges to put Claude tools directly into coding curricula. As Claude Code becomes a standard part of developer education, you may find more collaborators and junior hires already familiar with the workflow.

Wall Street Journal · 4 min read

Exclusive | Anthropic Takes Big Step in AI Race to Reshape College Coding Courses

Anthropic has partnered with hundreds of community and state colleges to put Claude tools directly into coding curricula. As Claude Code becomes a standard part of developer education, you may find more collaborators and junior hires already familiar with the workflow.

BBC News · 3 min read

Anthropic AI safety researcher quits with 'world in peril' warning

A senior Anthropic safety researcher resigned this week with a public warning about AI risk, drawing attention to internal tensions at the company as it scales rapidly. The departure is a notable data point for anyone tracking how Anthropic balances growth with its safety mission.

BBC News · 3 min read

Anthropic AI safety researcher quits with 'world in peril' warning

A senior Anthropic safety researcher resigned this week with a public warning about AI risk, drawing attention to internal tensions at the company as it scales rapidly. The departure is a notable data point for anyone tracking how Anthropic balances growth with its safety mission.